Collisions of neutral molecules and clusters is the most prevalent pathway in atmospheric new particle formation (NPF), with direct implications on air quality and climate. Until recently, these collisions have been modeled mainly using non-interacting hard sphere (NHS) models, which systematically underestimate collision and particle formation rates due to omission of long-range interactions. Lately, atomistic simulations which account for long-range interactions have been used to study neutral molecule-molecule and molecule-cluster collisions, but studies on cluster-cluster collisions have still been lacking despite the relevant role they can play e.g. in haze formation in polluted urban areas. We have therefore studied collisions between neutral clusters of <em>N</em> bisulphate and <em>N</em> dimethylammonium ions at <em>T </em>= 300 K up to <em>N </em>= 32 using atomistic molecular dynamics (MD) simulations. Direct simulation results have then been compared against both the traditional NHS model and the newly proposed interacting hard sphere (IHS) variant. We find the collision rates given by the NHS to be enhanced by a factor of 2.18–5.61 in the atomistic MD simulations, with enhancement decreasing with cluster size, and an asymptotic limit ≈ 2. The IHS model yields a constant enhancement factor of 3.36 over the NHS model for collisions between same-sized clusters, which decreases with increasing cluster size ratio. Our results demonstrate how even collisions between clusters of tens of acid-base pairs at a relatively high temperature cannot be accurately modeled by neglecting long-range interactions. We also show that the MD results cannot be reproduced by simple point-particle models, highlighting the importance of atomistic details of intermolecular interactions.